In Phaedrus, Plato quotes a story in which the god of writing appears to an early Pharaoh holding his new invention, the hieroglyphic script. The king tells the god to take it away, because it would ruin his subjects’ powers of memory and concentration, and fill them with the delusion that they knew things when they did not. The Pharaoh may have had a point, but Plato was not above being perverse, especially when it came to popular culture. It is also clear that he intended the anecdote as an anecdote, not as a programme of reform.
It is a commonplace that writing began through the use of pictures. The earliest examples come from southern Mesopotamia, and date to the end of the fourth millennium BC. They were used to record economic transactions, and there remained a distinctly mercantile quality to Mesopotamian civilisation throughout its history. Egypt soon followed, with similar ideas but with differing emphasis, since Egyptian writing seems to have been intended from the outset as a record of the royal court, its achievements as well as its bureaucracy. Both these scripts were pictorial, although the medium of Mesopotamian writing (a reed stylus impressed into clay tablets) meant that the original pictures were soon obscured. Egypt used pen and ink on papyrus, and also developed cursive forms of the script, but the hieroglyphs were kept for religious and decorative purposes, and lasted three and a half millennia until the arrival of Christianity.
Pictures are good for drawing visible objects, but what about ideas that cannot be portrayed in this way? And an even worse problem is posed by abstractions such as grammatical endings, which do jobs that may be conceptualised but can hardly be put into words, let alone drawn. One solution which appeared in the Near East was the method of the pun or charade: pictures were allowed to stand for other words with different meanings but similar sound. In heraldry, the same principle is known as canting: the name ‘Burton’ can be represented as a thistle (burr) accompanied by a barrel (tun). By this method, belief, a concept difficult to draw except by elaborate analogies, can be represented by pictures of a bee and a leaf, and so on. In recent years, progress has been made in the decipherment of the Maya script, and this, too, has turned out to be based on pictorial puns. American writing is unlikely to have been influenced by any civilisation in the Old World, so the conclusion seems to be that the phonetic principle is a universal of the human mind. The logical outcome of this process is the use of signs largely, or even solely, for their phonetic values: the letters of our alphabet are merely the extension of this principle to its extreme.
Another principle, common in Chinese writing, and sometimes found elsewhere, is that of the cartoon, where pictures are arranged in a combination suggesting associated ideas. In the Chinese script, the word ‘reliability’ can be expressed by the sign for ‘man’ accompanied by that for ‘speech’, since reliability can be defined as a man standing by his word. This concept is independent of the sounds of the signs used, and one needs to know Chinese to be able to pronounce it, but the method is a helpful one, within limits. The use of idea-signs, phonetic signs and various combinations of the two, means that early writing systems tend to be complex, with as many as one or two thousand signs, or more in the case of Chinese. They lead also to polyvalency, i.e. using the same sign in more than one function. Phonetic scripts (syllabaries and alphabets) appeared later, in the peripheral areas of the eastern Mediterranean and the Far East, by a process of abstraction which was much the same in both parts of the world. Nevertheless, elements of syllabic and alphabetic writing are present from the beginning in complex scripts, and have a habit of attaching themselves to the grammatical endings and particles which are ideographically hardest to define. Phonetic scripts are simpler, rarely exceeding a hundred signs, and coming down to as few as 22 in the case of the Phoenician alphabet which passed to Greece.
The outstanding undeciphered script is the writing of the Indus Valley civilisation of Pakistan, which is now the subject of a major study by Asko Parpola. This attractive and sophisticated writing is known chiefly from the magnificent seal impressions found in the ruins of Harappa, Mohenjo-Daro and Chanhu-Daro, although economic contacts spread them as far west as Mesopotamia. They date from the late third millennium BC, perhaps before the arrival of the Indo-European element central to later Indian civilisation. Several attempts have been made to decipher this tantalising script, but the problem is the limited nature of the texts, which are mostly short captions, and the lack of external controls. Nobody knows what language the Harappans spoke, and their names seem to be absent from later Sanskrit tradition. A Babylonian list of merchants from the Indus Valley complete with their names would be a godsend, but none has so far been found. A simple alphabet can often be cracked by the code-breaking techniques of counting the frequency of the particular signs and analysing the combinations in which they occur. If the underlying language is known, the process is normally straightforward. The Mayan script is more complex, similar to Egyptian hieroglyphs or Mesopotamian cuneiform, and the individual glyphs have a baroque quality to them which makes identification fearsomely difficult, but the underlying language was not seriously in doubt: dialects of Maya are spoken in the same areas to this day, and the local Indians have strong traditions which link them with their ancestors.
By contrast, Michael Ventris’s decipherment of the Cretan writing known as Linear B, an unknown language written in an unknown script, was an act of genius. Unlike Champollion and most other decipherers, Ventris had no bilingual text to guide him. The Indus Valley presents similar problems, since there are no good traditions in later sources, and no bilingual source has so far been identified. But the lack of hard evidence, and the potentially depressing nature of the task, has never deterred would-be decipherers. A celebrated example is the Phaestos disk, a unique piece of writing found in Crete, remarkable in that its signs are impressed into the clay of the disk: a process which anticipates printing, using a set of metal stamps or similar device. Batty decipherments of the Phaestos disk are not unknown; they tend to acquire a semi-mystical tone, making the object out to be a hymn to the ‘milk of the obsidian he-eagle’, or some other concept which testifies to the decipherer’s fecundity of mind but does nothing for his credibility. The arcane appearance of the disk attracts this sort of thing, although the odds must be in favour of the text turning out to be routine.
Parpola is aware that similar problems beset the Indus Valley inscriptions, and starts with methodological rigour: a count of the available signs suggests a complex script, similar to that of Egypt and Mesopotamia, and close analysis of particular signs reveals likely variants and common combinations. So far so Bletchley Park. Parpola then argues that the language of the Indus Valley was a form of Dravidian, a family of languages now spoken mainly in Southern India, but which isolated pockets suggest may have been spoken all over the sub-continent before the arrival of Indo-European. Common sense suggests that Dravidian is a strong candidate, but it is not the only one, since the linguistic map of India is extremely complex. Next he postulates that one of the most common Harappan signs, which resembles a fish, is the word for ‘star’, since the words for ‘fish’ and ‘star’ are identical, or nearly so, in later Dravidian. This gives him the key to a world of astrology, religion and folklore which survives in Hindu mythology to this day. The problem is that this method of decipherment seems to contradict the phonetic principle: in other writing systems, pictorial signs are normally extended, by phonetic resemblance, to words which cannot be expressed pictorially. Stars are easy to draw, so why should they be represented by pictures of fish, which are harder? It may be that there were overwhelming religious objections in the Indus Valley to drawing stars, or equally strong attachments to fish, but these need to be demonstrated. The rich survey of folklore which Parpola offers does not quite do this. Another weakness is that he doesn’t take his decipherment very far. This is virtuous reticence, but if he could demonstrate that at least one text can be plausibly read by his method, many false starts and contingent errors would surely be forgiven him, as they are forgiven any pioneer. A pity, since he may well be on the right lines, both about the nature of the script and even about the underlying language. He is thinking about deciphering, and in a stimulating way, but this is not decipherment. Instead, what Parpola has done is to argue for the unity of Indian civilisation over five millennia, in religion, folklore and iconography; which is perhaps more important than a decipherment.
Hieroglyphs are ancient Egypt, rather as a person’s name seems to contain his or her essence. The Chinese attachment to their complex writing system is not inertia, nor is it purely practical, forced on them by the existence of separate languages within China. It is to do with what it means to be Chinese. The Greek Septuagint and the Latin Vulgate are forms of the same text, but they are by no means the same book. Indeed, it is one of the paradoxical features of writing that the medium which enshrines a culture can also be used to subvert a culture. William Tyndale, the translator of the Bible into English vernacular, knew this well, and so did his opponents. The issue which divided them was real. Tyndale was threatening to unleash a force that might unite society, but which might equally tear it apart. Do we possess writing or does it possess us?
David Olson’s The World on Paper explores the relationship between thought and writing from a different perspective. Olson is a cognitive scientist, and a brave man. ‘Contrary to writers from Aristotle to Saussure,’ he tells us towards the end of the book, ‘I have argued that writing is not transcription of speech but rather provides a model for speech; we introspect language in terms laid down by our scripts.’ In other words, writing determines language and imagery, in much the same way that maps determine travel. This is reminiscent of the so-called Sapir-Whorf hypothesis, which holds that language itself lays down the pattern for thought. Carried to extremes, this would mean that languages exist which cannot be translated into any other. Few authorities, if any, can be found who agree with this. Language and thinking probably form a symbiosis, and it is possible that the nature of our writing system does colour the way we classify the world; but this may be all that it does. The Chinese may see certain problems differently from the Norwegians, but the differing nature of their scripts can be only one factor in the divergence. The same may be true of illiteracy as opposed to literacy: if the difference between the two were unbridgeable, monk and peasant would have developed into different species. On Olson’s own analogy of mapmaking his theory might mean that nobody has yet got round to exploring anything, since we are still waiting for somebody to come up with the first map of the terrain. To recall another ancient Egyptian story, written language is both the sorcerer and the apprentice.